Asymmetric PCR increases efficiency of melting peak analysis on the LightCycler
Aniko´ Szilva´si, Hajnalka Andrikovics, Lajos Kalma´r, Andra´s Bors, Attila Tordai*
Laboratory of Molecular Genetics, Institute of Hematology and Immunology, National Medical Center, Dio´szegi u´t 64, Budapest H-1113, Hungary Received 22 July 2004; received in revised form 1 April 2005; accepted 8 April 2005
Available online 27 June 2005
Abstract
Objectives: To systematically analyze the effects of asymmetric PCR on LightCycler melting analyses of four different allelic-discrimination systems and to reduce an inconsistent non-specific melting peak observed during factor V Leiden genotyping.
Design and methods: PCR amplifications and melting analyses were carried out with various oligonucleotide concentrations and ratios.
To monitor the efficiency, calculated peak area values were compared after melting analyses.
Results: Peak area values increased by a mean of 11.2-fold (range: 6 to 17) in case of an amplification primer ratio of 1:6.7 asymmetric PCR compared to symmetric primer conditions in four different SNP-genotyping systems. Using a complementary hybridization probe set for factor V Leiden genotyping, a converse amplification primer ratio was necessary for similar results.
Conclusions: Asymmetric PCR resulting in the formation of higher amounts of target strands significantly increases the efficiency of LightCycler allelic-discrimination.
D2005 The Canadian Society of Clinical Chemists. All rights reserved.
Keywords:Asymmetric PCR; LightCycler; Allelic-discrimination; Hybridization probes
Introduction
As the number of known causative single base mutations (single nucleotide polymorphism, SNP) increases, there is an increasing need for adequate, robust, and reliable techniques for SNP-genotyping[1]. Real time PCR ampli-fication followed by fluorescence-labeled probe hybrid-ization and melting curve analysis by the LightCycler is a frequently used single-tube technique [1]. Among other optimization parameters, the relative amounts of amplifica-tion primers are important, and a few reports menamplifica-tion the beneficial effects of asymmetric PCR (AS-PCR[2 – 4]).
Upon genotyping for the R506Q mutation of the factor V gene (FV Leiden) with in-house designed oligonucleotides,
an inconsistent third peak and various melting peak heights were observed. Systematic analyses of various oligonucleo-tide conditions of an AS-PCR were carried out by which signal intensity, specificity, and reproducibility improved significantly. The effects of using hybridization probes annealing to the complementary strand compared to the original setting were also examined. To extend our observations to additional gene- and SNP-environments, the characterization of the following SNP detection systems was also carried out: FII g.20210G>A (prothrombin); HFE (hereditary hemochromatosis) H63D and BCRP (breast cancer resistance protein or ABCG2) Q141K[5 – 8].
Methods Primer design
Amplification primers and hybridization probes were designed by the LightCycler Probe Design software
Abbreviations: AS-PCR, asymmetric PCR; BP, basepairs; FRET, fluorescence resonance energy transfer; PA, peak area; PCR, polymerase chain reaction; SNP, single nucleotide polymorphism; TM, melting temperature.
* Corresponding author. Fax: +36 1 3852255.
Clinical Biochemistry 38 (2005) 727 – 730
(Roche Diagnostics), and all oligonucleotides were synthe-sized by Integrated DNA Technologies (Coralville, USA).
Amplification primers Leiden-LCF and Leiden-LCR rec-ognizing the factor V gene were used, yielding a PCR product of 210 bp encompassing the c.1691G>A (R506Q) FV Leiden mutation. Two probe sets hybridizing to the amplification product at the same internal sequence but to the different strands were also designed. The ‘‘sense’’
probe set consists of Leiden-3LC and Leiden-5LC, and the
‘‘antisense’’ probe set consists of Leiden-3LC-comp and Leiden-5LC-comp. Leiden-5LC is perfectly matching with the wild type allele (c.1691G) of the sense strand (formed by Leiden-LCR amplification primer), while Leiden-5LC-comp is perfectly matching with the mutant allele (c.1691T) of the antisense strand (formed by Leiden-LCF amplification primer). These probe sets are complementary to each other, with one mismatch at the position of the Leiden mutation. Oligonucleotide sequences (FV Leiden, FII g.20210G>A, HFE H63D, BCRP Q141K) are available upon request.
PCR amplification and melting curve analyses
All amplification reactions were performed in a Light-Cycler. PCR was carried out in a final volume of 20AL and contained 10AL 2PCR Master Mix (Promega). Reactions were supplemented with 0.7 U Taq DNA polymerase (Finnzyme, Espoo, Finnland) and 0.75 mM MgCl2. Differ-ent amounts of amplification primers, 0.25Amol/L of each hybridization probes, and 200 ng DNA template were added. DNA was denatured at 95-C for 30 s and amplified through 70 cycles of PCR (95-C 0 s, 50-C 10 s, and 72-C 10 s [15 s for BCRP Q141K]).
Melting curve analyses were carried out by lowering the temperature to 40-C for 60 s and gradually increasing it to 80-C, with a transition rate of 0.1-C/s, and fluorescent signal was measured continuously.
The LightCycler software automatically generated melt-ing curve graphs and calculated the meltmelt-ing peak area (PA) values. PA calculations were used to compare efficiencies of different PCR conditions. Statistical analyses were carried out with the Student’s t test, with a level of significance of P < 0.05.
Results
After optimizing our in-house allelic-discrimination based on the LightCycler system for FV Leiden genotyping, we were able to consistently distinguish wild type and mutant alleles, but we observed an additional internal peak (at lower temperature) for some of the wild/wild genotype samples. By using an altered amplification primer ratio of 0.15:0.50Amol/L (1:3.3) forward:reverse (Leiden-LCF:Lei-den-LCR), signal intensities substantially increased, and the
We used FV Leiden genotyping as a model to system-atically analyze the effects of amplification primer ratio alterations. To this end, we carried out a series of LightCycler PCR reactions using altered amplification primer ratios with otherwise identical conditions. As an efficiency parameter, we compared the peak area (PA) values automatically generated by the LightCycler software.
By using the reverse amplification primer in excess compared to the amount of the forward primer, gradual increases in PA values were observed (Fig. 1). In the case of the 1:3.3 ratio, a 4.7-fold PA value increase (0.13T0.049 vs.
0.028 T 0.017, SD values are based on three parallel experiments) was observed compared to the 1:1 primer ratio (P= 0.014). At the ratio of 1:6.7, a further PA value increase of 2.4-fold (0.31 T 0.12 vs. 0.13 T 0.049) was measured compared to the 1:3.3 condition (P= 0.035). Further relative increases in the amount of the reverse primer (ratio of 1:13 and 1:16) caused some but non-significant PA value increases suggesting a plateau-effect (Fig. 1). Alteration of the primer ratio in excess of 1:16 did not result in further PA value increases (not shown). We also tested the effects of parallel increases of both primer concentrations (up to 0.5 Amol/L) without altering the 1:1 ratio, as well as increasing the amounts of the hybridization probes (up to 0.5Amol/L), and in all conditions, the PA values showed non-significant, inconsistent fluctuations (not shown).
By designing an alternative probe set, we used a novel approach to confirm the previously suggested hypothesis
Fig. 1. Changes in peak area (PA) values as a function of different amplification primer ratios in four different SNP genotyping systems. PCR reactions for the detection of FV Leiden (Leiden), FII g.20210G>A (FII);
HFE H63D (H63D) and BCRP Q141K (Q141K) sequence alterations were carried out with identical conditions including a homozygous wild/wild (except H63D) DNA template (see the Methods section), except for different forward:reverse amplification primer ratios. These ratios are indicated on the Fx_ axis with absolute numbers in Amol/L final concentrations with relative ratios in parentheses. These discrete separate conditions represent monotonous increases in the relative amount of one primer. Peak area (PA) values for each determination were automatically calculated by the LightCycler software. For better comparison, we calculated relative PA increases by dividing the appropriate PA values with that of 1:1 primer ratio amplification. These latter PA values were the following, as calculated by the LightCycler software for each SNP: Leiden:
[2 – 4]that a preferential unbalanced formation of one strand of PCR product enhances the efficiency of allelic-discrim-ination. This alternative probe set (termed ‘‘antisense’’) is strictly complementary to the ‘‘sense’’ probe set except for one base, namely a G>A change in position 1691 in the sensor probe. Thus, this probe shows a stronger hybrid-ization to the FV Leiden allele compared to the wild type allele. As expected, to get an outstanding high signal, excess amount of the forward primer (Leiden-LCF) had to be used with the ‘‘antisense’’ probes (Fig. 2, trace 2). Since the sensor probe of the ‘‘antisense’’ probe set is complementary to the Leiden allele, in case of the wild type sample, we recorded a melting peak at a lower temperature compared to the similarly efficient ‘‘sense’’ probes and excess reverse primer combination (trace 1). The opposite combinations were totally uninformative as to genotyping (see the entirely overlapping traces 5 and 6). In cases of the usual conditions, i.e. symmetric amplification primer ratios, strongly decreased, but similar efficiencies were observed (traces 3 and 4). Omission of traces 1 and 2 allowing the magnifi-cation of these latter traces (by a different Fy_ scale) indicated the correct genotypes (not shown).
We carried out AS-PCR with the excess amount of amplification primers giving rise to the strand complemen-tary to the hybridization probes in three further genotyping systems (FII g.20210G>A, HFE H63D, BCRP Q141K). As shown inFig. 1, similar trends were observed in each case.
The relative PA values showed gradual increases as a consequence of increasing the amounts of the appropriate
primers. Peak area values increased by a mean of 11.2-fold (range: 6 to 17) in case of an amplification primer ratio of 1:6.7 AS-PCR compared to symmetric primer conditions in four different SNP-genotyping systems.
Discussion
Application of AS-PCR in the context of a hybridization system was first described by Lay et al.[9]. In this system, the forward amplification primer was labeled by Cy5 which participates in the FRET reaction with a single hybridization probe. In this setting, it was obvious that the labeled primer should be used in excess.
In dual probe systems, i.e. when unlabeled amplification primers and two hybridization probes are used, the significance of AS-PCR was first indicated by Bernard et al. [10], who modeled the competition between the complementary strand of the PCR with the labeled probes by adding synthetic competitor oligonucleotides. As a result, an approximately 40% decrease in fluorescence signal intensity was observed. Burggraf et al. [3] reported incorrect genotyping results, observed during an in-house developed assay. Gradually decreasing the amount of the reverse primer ratio to 5:1, a gradual appearance of the second peak was demonstrated allowing the correct geno-typing of the heterozygous sample. In contrast to our observation, no significant increase in PA values occurred in this system as a consequence of AS-PCR.
Fig. 2. Comparison of melting curve analyses using sense and antisense hybridization probe sets and different amplification primer ratios. PCR was carried out byFsymmetric_(0.25Amol/L of each primer; traces 3 and 4) andFasymmetric_PCR (primer ratios of 0.15:0.5Amol/L, i.e. 1:3.3; traces 1, 2, 5, and 6) as described in the Methods section, with the excess of both Leiden-LCR (reverse, exR) or Leiden-LCF (forward, exF) primers and with both probe sets (sense probes: Leiden-5LC and 3LC; antisense probes: Leiden-5LC-comp and 3LC-comp). Melting curve analyses were carried out, and negative derivates of changes in fluorescence intensities are illustrated as a function of temperature as automatically calculated by the LightCycler software. The highest melting peak was observed if the sense probe set was used in the presence of an excess of the reverse primer (trace 1). A smaller melting peak was recorded if the antisense probe set was used with an excess of the forward primer (trace 2). However, the opposite ‘‘unfavorable’’ primer and probe set combinations (traces 5 and 6) resulted in
Barratt et al. [4] observed a ‘‘hook effect’’ in a herpes simplex virus genotyping system. This phenomenon can decrease fluorescence signal in the latest cycles and can also compromise genotyping and affect consistency during melting analysis. All problems caused by the ‘‘hook effect’’
were eliminated by AS-PCR, however, no details on exact concentrations and consequences were described. We also observed low signal intensities but without a ‘‘hook effect’’
(not shown).
None of the above studies focused specifically on the role of AS-PCR and approached this issue by designing an alternative probe set complementary to the opposite strand.
Applying the complementary hybridization probe set in our FV Leiden model system gave the predicted result, namely, the opposite amplification primer ratio was necessary to achieve a similar response compared to the original probe set and amplification primer ratio (Fig. 2).
We carried out AS-PCR with different amplification primer ratios in four in-house designed SNP detection systems, namely, FV Leiden, FII g.20210G>A, HFE H63D and BCRP Q141K (Fig. 1), and we observed similar trends in each system, indicating that the advantages of altered amplification primer ratios favoring the strand complementary to the hybridization probes are likely to be valid in several gene-environments.
In summary, our studies have demonstrated that (i) AS-PCR using excess amounts of amplification primers allowing the preferential synthesis of the strand comple-mentary to the hybridization probes results in significant increases in fluorescence signal intensity and improves specificity during melting analyses; (ii) similar trends can be observed in each investigated SNP genotyping systems;
(iii) this phenomenon can also be reproduced with opposite altered ratios if identical amplification primers and PCR conditions are used with novel hybridization probes annealing to the complementary strand; (iv) testing the role of altered amplification primer ratios should be an integral part of the optimization procedure of a newly developed in-house dual probe allelic-discrimination
sys-tem, especially if signal intensities are low or non-specific peaks are observed.
Acknowledgments
The authors thank Pfundt Antalne´, Bakonyi Ildiko´, Horva´th Csongorne´, and Szaver Gabriella for technical support. This work was partly supported by a grant from OTKA T034830. A.T. is a recipient of the ‘‘Bolyai Ja´nos’’
fellowship.
References
[1] Lyon E. Mutation detection using fluorescent hybridization probes and melting curve analysis. Expert Rev Mol Diagn 2001;1:92 – 101.
[2] Lyon E, Millson A, Phan T, Wittwer CT. Detection and identification of base alterations within the region of factor V Leiden by fluorescent melting curves. Mol Diagn 1998;3:203 – 9.
[3] Burggraf S, Ko¨sel S, Lohmann S, Beck R, Olgemo¨ller B. Unexplained DNA melting behavior in a genotyping assay. Clin Chem 2002;
48:199 – 201.
[4] Barratt K, Mackay JF. Improving real-time PCR genotyping assays by asymmetric amplification. J Clin Microbiol 2002;40:1571 – 2.
[5] von Ahsen N, Schutz E, Armstrong VW, Oellerich M. Rapid detection of prothrombotic mutations of prothrombin (G20210A), factor V (G1691A), and methylenetetrahydrofolate reductase (C677T) by real-time fluorescence PCR with the LightCycler. Clin Chem 1999;45:694 – 6.
[6] Bernard PS, Ajioka RS, Kushner JP, Wittwer CT. Homogeneous multiplex genotyping of hemochromatosis mutations with fluorescent hybridization probes. Am J Pathol 1998;153:1055 – 61.
[7] Tag CG, Gressner AM, Weiskirchen R. An unusual melting curve profile in LightCycler multiplex genotyping of the hemochromatosis H63D/C282Y gene mutations. Clin Biochem 2001;34:511 – 5.
[8] Honjo Y, Morisaki K, Huff LM, et al. Single-nucleotide poly-morphism (SNP) analysis in the ABC half-transporter ABCG2 (MXR/BCRP/ABCP1). Cancer Biol Ther 2002;1:696 – 702.
[9] Lay ML, Wittwer CT. Real-time fluorescence genotyping of factor V Leiden during rapid-cycle PCR. Clin Chem 1997;43:2262 – 7.
[10] Bernard PS, Pritham GH, Wittwer CT. Color multiplexing hybridi-zation probes using the apolipoprotein E locus as a model system for genotyping. Anal Biochem 1999;273:221 – 8.